Head-to-head verdicts that say how sure we are — every recommendation is labeled strongly supported, weakly supported, disputed, or outdated.
We tested both on a 200k+ LOC codebase. Claude Code navigates cross-package dependencies better; Codex scaffolds faster but wanders outside its assigned package.
iOS is the acid test: Xcode builds, simulators, and API churn. With XcodeBuild MCP wired up, Claude Code pulls clearly ahead; without it, both are guessing.
New repo, empty directory, ship a working skeleton. Codex sprints, Cursor iterates, and the community is genuinely split on which habit you want at day zero.
Terminal agent vs editor agent on the refactors that touch 20+ files. Claude Code wins on follow-through; Cursor wins when you need to steer mid-flight.
Free tier vs metered Opus, measured on what actually shipped. Gemini's sticker price wins; the rework rate eats most of the difference. Small sample, weak conclusion.
Bring-your-own-model transparency against a tuned first-party pipeline. Early evidence says the polish gap is real but narrowing — and priced accordingly.